- Chatbots like Claude are designed to feel personal but are not private, with conversations visible to the company and potentially indexed by search engines.
- Claude share links were indexed by Google, exposing sensitive user chats including legal and personal discussions, though Anthropic quickly fixed the issue.
- The leak primarily affected shared chats, but some users who never shared their chats may also have been exposed.
- AI companies commonly use user chats for training and may have human reviewers, and data can be retained even after deletion.
- Users should never share personal or sensitive information with AI chatbots to protect their privacy.